v0.4.0 - swarm detection and the exec hand-off

A name catching fire is visible in the data before anyone knows which token is
going to be the one. This release catches that moment and answers the question
it raises, which is not "will this pump" but "eleven of these are copies, which
one is not".

Added

• novamp swarm. A burst of launches collapsing into one cluster inside a
  short window, from wallets that are not each other, is a crowd reacting to
  something. The distinct-deployer bar is what keeps one farm spraying a ticker
  from ten wallets out of the results.
• News correlation with the order checked. A match requires the word, a
  timestamp, and publication before the first launch. An undated item is
  dropped rather than assumed fresh; a headline stamped after the burst is
  labelled rather than quietly counted.
• Public Telegram channels as a source, read through the t.me/s/ web
  preview. No bot token, no API id, nothing joined, no history walk.
• The selection rule in one file. A copy can never win on points; the gates
  that follow are absolute rather than relative, so NO PICK is a real answer
  and is most of them.
• The hand-off. swarm --exec runs a program you wrote and gives it the
  finding as JSON on stdin. novamp still holds no key and signs nothing, and the
  no-signer CI job now scans examples/ as well. See docs/EXEC.md.
• 50 new tests: 177 total.

Fixed

• The densest-window search preferred the later window on a tie, which cut the
  first born out of its own swarm. Ties now go to the earlier window, and
  ranking runs over the burst plus everything that launched before it.
• The phrase check compared a tight-folded term against a loose-folded headline,
  so a multi-word token name never matched a headline containing it.
• channelUrl did not strip a bare t.me/name without a scheme.

Full changelog: CHANGELOG.md

novamp 0.3.0

novamp finds every token fighting over the same name on Robinhood Chain, works out which one is the original, and flags the ones to leave alone.

A name catches, and sixty seconds later there are fourteen tokens carrying it. Most of what retail loses on this chain is not lost to the rug they bought, it is lost to buying number seven while reading number one's timeline.

Read only. No key, no signer, no transaction path, and CI fails the build if a signing primitive ever appears in the source.

This is the first tagged release. Everything below is in it.

What it does

Paste a ticker or a name. Every launch under it gets a label:

• ORIGINAL — first, and nothing against it
• TAINTED — first, but the operator loaded it himself
• CONTESTED — not first, but the money is here anyway
• VAMP — a later copy with nothing of its own
• DEAD — nothing is trading

It matches on the ticker and the token name, joined transitively, folding lookalike characters, zero width characters, leetspeak, plurals and filler words. A copy that takes a fresh symbol and keeps the name does not slip past.

Being first is not the same as being real. A first launch carrying a heavy dev buy, a stack of wallets exempt from the opening tax, and a deployer with a hundred launches and no graduations is called TAINTED, and novamp then shows where the flow actually went.

What it reads

• the cluster: who was first, how far behind each copy landed, top-10 concentration with the bonding curve and the pool excluded, dev share, wallets exempt from the opening tax, deployer history
• the sell side: whether the deployer sold and how fast, whether one sale took a bite out of the reserve it hit, whether more quote is leaving the curve than entering it
• the early money in quote, not in wallets: total in, median ticket, and the share held by the three biggest buyers. Counting addresses is exactly what a farm is optimised against, because addresses are free
• proven wallets on realized results: quote out against quote in, per position. A position with no sale is open, not a win
• your own wallet: which of your positions turned out to be copies, with the original beside each one
• the operators: deployers ranked by copies shipped and by how many other wallets share their funder

Every point of the 0-100 score prints its own reason next to it, so you argue with a line rather than with a number.

Commands

vamp scan wallet farms board watch watchlist stats narrative index smart doctor

All of them take --demo and run on the bundled fixtures with no RPC, no config and no key. board puts the same engine behind a page on 127.0.0.1.

Honest about what it is not

The score orders a cluster. It does not predict a price, and the README says so rather than hiding it.

The chain readers are written against the documented pons v2 ABI and have not yet been exercised against mainnet by the author. The reserve behind a sale is reconstructed from buy and sell events rather than read, and only a wallet's first buy on a launch is counted. All of it is written down in docs/LIMITATIONS.md, along with everything else that is approximate.

127 offline tests. MIT. Runs on your machine, not anyone else's.